Silicon is a critical component in the radio access network (RAN) business. Just ask Nokia. The company made some mistakes in silicon in the early days of 5G and is still trying to dig itself out of the problems caused by those ill-advised decisions.

“Nokia ended up in a situation some three years ago where we had to commercialize our first 5G products without any system-on-a-chip (SoC) for massive MIMO (multiple input, multiple output) 5G layer 1, so the physical layer, and eCPRI (enhanced common public radio interface),” Tommi Uitto, president of Nokia’s Mobile Networks business, told SDxCentral in a recent phone interview.

“We didn’t have the right SoC for this 25 Gb/s interface. Now that’s all history because I have the first generation of SoC for all our base station functions, and we have the next generation already in the works,” he said.

Indeed, Nokia last year began shipping newer 5G RAN equipment, which features a more competitive and feature-rich SoC. Those products accounted for 44% of all shipments during its most recent quarter, and the company said it’s on track to hit its target of 70% by the end of 2021. It expects all 5G shipments to be equipped with the new SoC by the end of next year.

“Silicon is a very, very important part in the overall competitiveness because it really [defines] the ability to support certain features and functionalities: the connectivity, the capacity, power consumption, product size and weight, and product cost. All of these are important things in the radio business,” Uitto said. 

“Hindsight is perfect science because you can only be right, but I think Nokia probably overestimated the need for having the flexibility for FPGA when the 3GPP standard was still evolving,” he said. “And I think Nokia also did not properly see how non-standalone 5G was going to be introduced before standalone.”

Both of those developments “caught Nokia a little bit by surprise,” but “that’s really a long time ago,” Uitto added. “Today we have the SoCs, and I’m super happy about what we have in our SoC research and development (R&D) today.”

Those misfires on silicon precipitated a turnaround effort in Nokia’s RAN business in 2019, and that largely consisted of “making improvements in how we develop the products, the products themselves, the product platform, SoC, the baseband platforms, the radio platforms,” he said. 

The full benefit of those efforts haven’t been realized yet because Nokia continues to introduce new products and more are coming later this year. Massive MIMO 5G radios that Nokia previously revealed will also be released “soon,” and more specifically in time to compete head-to-head with competing massive MIMO products, CEO Pekka Lundmark said this week on Nokia’s earnings call. 

Lundmark also claimed Nokia’s Massive MIMO 5G equipment will have wider bandwidth and weigh less than the competition.

“This is really the sort of comeback year in 5G,” Uitto said. “And, of course, it’s not just about the catch-up, but then going for leadership in selected areas in 5G.”

Nokia’s RAN boss said he’s increased 5G R&D with a 40% increase in headcount. “The output has grown by about 60%, so output has grown faster than the input, which means that productivity has improved,” he said. 

“In some areas of R&D I have grown by more than 100% over the past two years in 5G, and especially in SoC development, even by 200%, so I have tripled our SoC R&D,” Uitto said. 

Lundmark promoted Uitto to run the reformed Mobile Networks business unit as of Jan. 1, and effectively combined the RAN products business with related services, including network planning, deployment, and technical support. 

“In the past, Nokia had these separate and we can generate value out of this combination because now things like design for services, optimizing the total site cost, ease of deployment, they come much more naturally,” Uitto explained. “Now we’re running it in just one dimension.”